It's nothing new that Machine Learning and Big Data have been making the headlines for quite some time now, almost half a decade at least. The Data Scientist job has even been identified by Glassdoor as the
best job in America (as of April 2016).
However, according to Gartner's 2015 Hype Cycle for emerging technologies, Machine Learning has now passed the peak of inflated expectations and is now sliding down the trough of disillusionment, and so is Big Data (in Gartner's 2015 Hype Cycle for Business Intelligence and Analytics). Is Data Science just a fad? Are we starting to see the end of this job market bubble?
"Show me the
According to various job boards, not yet! Not for a while at least. Demand is exploding, Online trainings have been multiplying across many platform (Coursera, Udacity, edx to name only a few) to try to catch up with the increasing demand!
It makes perfect sense when considering the staggering amount of data generated every minute online (check Domo's famous infographic Data Never Sleeps 3.0) - Someone has to process it anyway, and who better than the data scientist, with a statistics background, some hacking skills and a sound business sense to gather, clean, analyze, and present his findings based on data (Please check Drew Conway's Venn Diagram on a more comprehensive definition of data science).
As Cuba Gooding Jr uttered the infamous "Show me the Money" mantra in Jerry Maguire, Business owners and stakeholders want to see data-based analyses to make data-enhanced decisions but it unfortunately takes more than a bit of excel tinkering to sift through terabytes of unstructured and unfiltered data.
The 5 Stages of Analytical Competition
I've been a business analyst since 2007, the year of the first iPhone, and when I asked what a business analyst was during my interview process, I was replied with something along the lines of a job at the cross roads of strategy, finance and retail (I would be working on the retail side with Expedia back then).
I really liked the swiss army knife type of approach this job would require (and which defined me well), but I also learned - thanks to Thomas H. Davenport and Jeanne G. Harris (Go read Competing on Analytics now if you haven't done so already)- that it takes more than a business analyst for a company to compete on analytics (it is a good start though), and that they come in five flavours with different objectives (this list is quoted from their book, chapter 2, p. 36):
- Analytically Impaired: getting accurate data to improve operations
- Localized Analytics: using analytics to improve one or more functional activities
- Analytical Aspirations: using analytics to improve a distinctive capability
- Analytical Companies: differentiating through analytics
- Analytical Competitors: fully competing on analytics
Since 2007, I can't say that I've encountered every category for each of my professional experiences, but I sure had a good representative sample of analytically driven and less analytically driven companies as employers. Both of them are very keen on leveraging data to improve performance (or whatever maximizes shareholder value), but the difference is mainly the speed at which you're able to spread the gospel when it comes to analytical good practices.
The Golden Triangle - Descriptive, Predictive & Prescriptive Analytics
As a company or a business unit, you belong in each of these categories based on the questions you're trying to answer.
- They can be Descriptive (for stage 1 companies): "What is exactly happening in our business?",
- Predictive (for stage 2 and 3 companies): "Why is this happening? What are the root causes? Will it keep on happening? or not?"
- Prescriptive (for stage 4 and 5): "how can we differentiate and innovate? What's next, how do we stay ahead?"
I don't really like the hierarchy implied in those stages, because it appears as if you're in stage 1, you're just a Report Monkey, aiming at becoming an Analytics Ninja in Stage 5. The truth is, that a good analysis often includes all three types of analytics:
- A compelling storytelling with a good description of the situation at hand (and the hype around data visualization certainly supports that) - the "What"
- A sound analysis on why this is happening - the "Why"
- And a course of action which will result in a business decision - the "How"
It shouldn't come therefore as a surprise that both Predictive and Prescriptive Analytics are on top of the Gartner's Analytics Hype Cycle peak. The Analytics market is slowly maturing and from the "Jack-of-all-trades" Data Scientists will emerge specialists to cover all the aspects of the vast Data Science field (remember the 1990's Webmaster?).
So, to wrap this article up, not only don't I think that Data Science is just a fad, but I'm a strong believer that Data has started to govern our lifes and actions, and with a strong sense of ethics, it will be infinitely beneficial.
- https://www.glassdoor.com/Best-Jobs-in-America-LST_KQ0,20.htm, Glassdoor: 25 best jobs in America in 2016
- https://www.gartner.com/doc/3106118, Gartner: Hype Cycle for Business Intelligence and Analytics, August 4th, 2015
- http://www.gartner.com/newsroom/id/3114217, Gartner's 2015 Hype Cycle for Emerging Technologies Identifies the Computing Innovations That Organizations Should Monitor, August 18th, 2016
- https://www.domo.com/blog/2015/08/data-never-sleeps-3-0/, Data Never Sleeps 3.0, August 13th, 2015
- http://drewconway.com/zia/2013/3/26/the-data-science-venn-diagram, Dew Conway's Data Science Venn Diagram
- Competing on Analytics, Thomas H. Davenport & Jeanne G. Harris, HBR Press, March 2007